f16c-intrinsics.ll
7.72 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avx,+f16c -show-mc-encoding -disable-peephole | FileCheck %s --check-prefixes=AVX,X86
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx,+f16c -show-mc-encoding -disable-peephole | FileCheck %s --check-prefixes=AVX,X64
; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avx512vl -show-mc-encoding -disable-peephole | FileCheck %s --check-prefixes=AVX512VL,X86-AVX512VL
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512vl -show-mc-encoding -disable-peephole | FileCheck %s --check-prefixes=AVX512VL,X64-AVX512VL
define <8 x i16> @test_x86_vcvtps2ph_128(<4 x float> %a0) {
; AVX-LABEL: test_x86_vcvtps2ph_128:
; AVX: # %bb.0:
; AVX-NEXT: vcvtps2ph $0, %xmm0, %xmm0 # encoding: [0xc4,0xe3,0x79,0x1d,0xc0,0x00]
; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3]
;
; AVX512VL-LABEL: test_x86_vcvtps2ph_128:
; AVX512VL: # %bb.0:
; AVX512VL-NEXT: vcvtps2ph $0, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0x79,0x1d,0xc0,0x00]
; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.vcvtps2ph.128(<4 x float> %a0, i32 0) ; <<8 x i16>> [#uses=1]
ret <8 x i16> %res
}
declare <8 x i16> @llvm.x86.vcvtps2ph.128(<4 x float>, i32) nounwind readonly
define <8 x i16> @test_x86_vcvtps2ph_256(<8 x float> %a0) {
; AVX-LABEL: test_x86_vcvtps2ph_256:
; AVX: # %bb.0:
; AVX-NEXT: vcvtps2ph $0, %ymm0, %xmm0 # encoding: [0xc4,0xe3,0x7d,0x1d,0xc0,0x00]
; AVX-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77]
; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3]
;
; AVX512VL-LABEL: test_x86_vcvtps2ph_256:
; AVX512VL: # %bb.0:
; AVX512VL-NEXT: vcvtps2ph $0, %ymm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0x7d,0x1d,0xc0,0x00]
; AVX512VL-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77]
; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.vcvtps2ph.256(<8 x float> %a0, i32 0) ; <<8 x i16>> [#uses=1]
ret <8 x i16> %res
}
declare <8 x i16> @llvm.x86.vcvtps2ph.256(<8 x float>, i32) nounwind readonly
define void @test_x86_vcvtps2ph_256_m(<8 x i16>* nocapture %d, <8 x float> %a) nounwind {
; X86-LABEL: test_x86_vcvtps2ph_256_m:
; X86: # %bb.0: # %entry
; X86-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT: vcvtps2ph $3, %ymm0, (%eax) # encoding: [0xc4,0xe3,0x7d,0x1d,0x00,0x03]
; X86-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_x86_vcvtps2ph_256_m:
; X64: # %bb.0: # %entry
; X64-NEXT: vcvtps2ph $3, %ymm0, (%rdi) # encoding: [0xc4,0xe3,0x7d,0x1d,0x07,0x03]
; X64-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77]
; X64-NEXT: retq # encoding: [0xc3]
;
; X86-AVX512VL-LABEL: test_x86_vcvtps2ph_256_m:
; X86-AVX512VL: # %bb.0: # %entry
; X86-AVX512VL-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-AVX512VL-NEXT: vcvtps2ph $3, %ymm0, (%eax) # EVEX TO VEX Compression encoding: [0xc4,0xe3,0x7d,0x1d,0x00,0x03]
; X86-AVX512VL-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77]
; X86-AVX512VL-NEXT: retl # encoding: [0xc3]
;
; X64-AVX512VL-LABEL: test_x86_vcvtps2ph_256_m:
; X64-AVX512VL: # %bb.0: # %entry
; X64-AVX512VL-NEXT: vcvtps2ph $3, %ymm0, (%rdi) # EVEX TO VEX Compression encoding: [0xc4,0xe3,0x7d,0x1d,0x07,0x03]
; X64-AVX512VL-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77]
; X64-AVX512VL-NEXT: retq # encoding: [0xc3]
entry:
%0 = tail call <8 x i16> @llvm.x86.vcvtps2ph.256(<8 x float> %a, i32 3)
store <8 x i16> %0, <8 x i16>* %d, align 16
ret void
}
define void @test_x86_vcvtps2ph_128_m(<4 x i16>* nocapture %d, <4 x float> %a) nounwind {
; X86-LABEL: test_x86_vcvtps2ph_128_m:
; X86: # %bb.0: # %entry
; X86-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT: vcvtps2ph $3, %xmm0, (%eax) # encoding: [0xc4,0xe3,0x79,0x1d,0x00,0x03]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_x86_vcvtps2ph_128_m:
; X64: # %bb.0: # %entry
; X64-NEXT: vcvtps2ph $3, %xmm0, (%rdi) # encoding: [0xc4,0xe3,0x79,0x1d,0x07,0x03]
; X64-NEXT: retq # encoding: [0xc3]
;
; X86-AVX512VL-LABEL: test_x86_vcvtps2ph_128_m:
; X86-AVX512VL: # %bb.0: # %entry
; X86-AVX512VL-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-AVX512VL-NEXT: vcvtps2ph $3, %xmm0, (%eax) # EVEX TO VEX Compression encoding: [0xc4,0xe3,0x79,0x1d,0x00,0x03]
; X86-AVX512VL-NEXT: retl # encoding: [0xc3]
;
; X64-AVX512VL-LABEL: test_x86_vcvtps2ph_128_m:
; X64-AVX512VL: # %bb.0: # %entry
; X64-AVX512VL-NEXT: vcvtps2ph $3, %xmm0, (%rdi) # EVEX TO VEX Compression encoding: [0xc4,0xe3,0x79,0x1d,0x07,0x03]
; X64-AVX512VL-NEXT: retq # encoding: [0xc3]
entry:
%0 = tail call <8 x i16> @llvm.x86.vcvtps2ph.128(<4 x float> %a, i32 3)
%1 = shufflevector <8 x i16> %0, <8 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
store <4 x i16> %1, <4 x i16>* %d, align 8
ret void
}
define void @test_x86_vcvtps2ph_128_m2(double* nocapture %hf4x16, <4 x float> %f4X86) #0 {
; X86-LABEL: test_x86_vcvtps2ph_128_m2:
; X86: # %bb.0: # %entry
; X86-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT: vcvtps2ph $3, %xmm0, (%eax) # encoding: [0xc4,0xe3,0x79,0x1d,0x00,0x03]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_x86_vcvtps2ph_128_m2:
; X64: # %bb.0: # %entry
; X64-NEXT: vcvtps2ph $3, %xmm0, (%rdi) # encoding: [0xc4,0xe3,0x79,0x1d,0x07,0x03]
; X64-NEXT: retq # encoding: [0xc3]
;
; X86-AVX512VL-LABEL: test_x86_vcvtps2ph_128_m2:
; X86-AVX512VL: # %bb.0: # %entry
; X86-AVX512VL-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-AVX512VL-NEXT: vcvtps2ph $3, %xmm0, (%eax) # EVEX TO VEX Compression encoding: [0xc4,0xe3,0x79,0x1d,0x00,0x03]
; X86-AVX512VL-NEXT: retl # encoding: [0xc3]
;
; X64-AVX512VL-LABEL: test_x86_vcvtps2ph_128_m2:
; X64-AVX512VL: # %bb.0: # %entry
; X64-AVX512VL-NEXT: vcvtps2ph $3, %xmm0, (%rdi) # EVEX TO VEX Compression encoding: [0xc4,0xe3,0x79,0x1d,0x07,0x03]
; X64-AVX512VL-NEXT: retq # encoding: [0xc3]
entry:
%0 = tail call <8 x i16> @llvm.x86.vcvtps2ph.128(<4 x float> %f4X86, i32 3)
%1 = bitcast <8 x i16> %0 to <2 x double>
%vecext = extractelement <2 x double> %1, i32 0
store double %vecext, double* %hf4x16, align 8
ret void
}
define void @test_x86_vcvtps2ph_128_m3(i64* nocapture %hf4x16, <4 x float> %f4X86) #0 {
; X86-LABEL: test_x86_vcvtps2ph_128_m3:
; X86: # %bb.0: # %entry
; X86-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT: vcvtps2ph $3, %xmm0, (%eax) # encoding: [0xc4,0xe3,0x79,0x1d,0x00,0x03]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_x86_vcvtps2ph_128_m3:
; X64: # %bb.0: # %entry
; X64-NEXT: vcvtps2ph $3, %xmm0, (%rdi) # encoding: [0xc4,0xe3,0x79,0x1d,0x07,0x03]
; X64-NEXT: retq # encoding: [0xc3]
;
; X86-AVX512VL-LABEL: test_x86_vcvtps2ph_128_m3:
; X86-AVX512VL: # %bb.0: # %entry
; X86-AVX512VL-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-AVX512VL-NEXT: vcvtps2ph $3, %xmm0, (%eax) # EVEX TO VEX Compression encoding: [0xc4,0xe3,0x79,0x1d,0x00,0x03]
; X86-AVX512VL-NEXT: retl # encoding: [0xc3]
;
; X64-AVX512VL-LABEL: test_x86_vcvtps2ph_128_m3:
; X64-AVX512VL: # %bb.0: # %entry
; X64-AVX512VL-NEXT: vcvtps2ph $3, %xmm0, (%rdi) # EVEX TO VEX Compression encoding: [0xc4,0xe3,0x79,0x1d,0x07,0x03]
; X64-AVX512VL-NEXT: retq # encoding: [0xc3]
entry:
%0 = tail call <8 x i16> @llvm.x86.vcvtps2ph.128(<4 x float> %f4X86, i32 3)
%1 = bitcast <8 x i16> %0 to <2 x i64>
%vecext = extractelement <2 x i64> %1, i32 0
store i64 %vecext, i64* %hf4x16, align 8
ret void
}